We are in the early stages of planning for an ASP solution that will service multiple sites and host multiple servers with a requirement that users attach to ONLY their terminal server in order to access the application and the connection to their data set. I would prefer that I NOT have to purchase multiple IP's to service every site connection so I am hoping that 2X will obviate that path. In a nutshell, the 2X server would face my router and serve as the intermediary to connect users to "their" server/data set. Each site will be configured on a blade server that will host 3 virtual machines running the database and the application as well as the Terminal Server for that "site". I would like to use the features of 2X to publish the applications to the end users (they will be serviced across either an MPLS or SSL/VPN using the Internet as the backbone into our colo facility) but I am unsure if 2X will provide any function in this scenario. Obviously, load balancing is not a part of this scenario as all I need for 2X to accomplish is to ensure that the proper users are connected to their proper terminal server. Has anyone had any experience in this type of configuration?
Hi there, 2X ApplicationServer & LoadBlancer does the job you mentioned. 1) Configure your blades with the vms. 2) Install Application Server & Loadbalncer. 3) Add all the VM Terminal servers in the TS page in the 2x Console 4) In the published apps page I suggest you create a group for each company and filter it either by AD Group or ips. 5) Publish the specific apps for that company in that group & configure all the apps in that group to run the app from a specific/ group of TS. 6) Clients need to connect to the 2X ApplicationServer & Loadbalancer. 7) The displayed apps will be filtered by the user name specified or IP Like this Users of a specific company will only make use of the configured TS Nixu
OK. Thanks for the reply. After more reading I suspected that it would work as you outlined. I plan to use a fairly robust box as the 2X server and each 2X server will front a single blade chassis. How well does 2X scale out? I anticipate that each blade chassis will be handling somewhere between 7-14 discrete companies with an average of about 35 users per company running a medical PM and EMR app. At what point does the ability of 2X begin to trend downwards?
Hi there, You can divide the required resources for a 2X ApplicationServer & Loadbalancer into the following: 1) Resources needed to monitor the Terminal Servers (Very low but constant). In your case you mentioned you are going to have 3 TS for each blade, so that should be negligible. 2) Resources need to loadbalance new requests (low but peaks up momentarily, since these occur on launching of a new application). One must only consider this resource requirement when one has a large number of published application and loads of filtering rules. 3) Resource needed to Proxy the connections. If clients connect in direct mode (clients can see the ip of the TS) will not use this type of resource. On the other hand SSL & Tunneled client connections make use of this resource. For tunneled connections, processor and memory resources are not an issue. Clients using SSL will make more use of these resources. What really effect scalability in this component is the network connection. USUALLY Network connections used ONLY for RDP scale up very easily with out any issues. On the other hand if the same network is heavily used for other protocols such as file sharing thus causing network latency, one can experience delayed responsiveness. Since every environment has it own variables it is very difficult to give any figures, so what i suggest is to install the system and monitor your resources, especially the network connection. Nixu
